I am looking for small size(preferably pcb mount) wifi antenna that is directional, more specifically half sphere coverage area. The product I build stays on the ground and I like utilized antenna that is always looking up.

I looked at several ceramic antenna datasheets but could not see one that fits. What type of antenna I should look at?


Based on the response, I thought I should define the problem differently. I have a working product where there is a wifi on top of the product and there are sensitive analogue circuits below the wifi antenna/module. I occasionally get spikes in my analogue circuit (analog has very high gain, about 5M times) when the module is receiving data (TX is fine). Currently, I slapped a EMI absorber on top of my analog stuff and shipped the product, unfortunately this also reduces my wifi coverage/signal by about 10db. In the next version of the product, I like to replace my omnidirectional antenna with a more directional one and avoid signal loss. Hence the question of perfect half sphere of antenna. Now that I know that is not possible, either I need an antenna that has a cone shape coverage or more shielding on Analog which I really like to avoid.


I would consider a patch antenna. With a thick enough substrate you can cover the entire Wi-Fi 2.4 GHz band. It has a pretty decent front to back ratio (>10 dB) and an almost onmi hemispherical pattern.

While it is impossible to say for sure from you description, it doesn't sound like the antenna is the problem. If it were, you would have had a lot more problems in Tx than Rx (Tx is usually at least 50 dB above Rx). The fact that you are having problems in Rx makes me suspect some sort of coupling between your LNA and analog circuitry. When you put absorbers, you are lowering the signal level in the LNA (or further down the receive chain) and decrease the coupling.

  • \$\begingroup\$ Yuriy how can I verify it is the lna coupling is the problem but not the antenna? I am not experienced in rf and I haven't been able to really understood the problem in detail. I trace the issue to wifi and end up using absorber to solve the issue. Without the absorber I see very sporadic behavior. For example if I move the cables inside the product, it would impact the noise figure I observe on analog. (Improve or reduce). \$\endgroup\$ – Ktc Jan 19 '14 at 17:18
  • \$\begingroup\$ Well, I can't say without getting more info on the issue so a pic/more data would definitely help. Would it be possible to remove the antenna and instead feed an RF signal of compatible amplitude? \$\endgroup\$ – Yuriy Jan 19 '14 at 18:49

A precise half sphere radiation pattern is pie in the sky realistically. The theoretical isotropic antenna transmits into free space a perfect spherical radiation but, it is only a theory used to help explain the Friis equations. They do not exist. Even if they did, and you could mount it on a groundplane/PCB, the radiation pattern along the axis of the groundplane would be zero. This is not half-spherical so you should consider your requirements and be prepared to compromise.

  • \$\begingroup\$ I would settle for large cone. I have several sensitive analog circuits below the wifi and omnidirectional antennas create load of headaches. So this time instead of fiddling with shielding etc. I like to use an antenna that doesn't disturb the analog stuff. \$\endgroup\$ – Ktc Jan 19 '14 at 11:35
  • \$\begingroup\$ What ones have you looked at so far - a link to a data sheet would be cool. \$\endgroup\$ – Andy aka Jan 19 '14 at 12:41
  • \$\begingroup\$ I have looked at fractus.com/sales_documents/FR05-S1-N-0-102/… There are several similar others. \$\endgroup\$ – Ktc Jan 19 '14 at 12:48
  • \$\begingroup\$ How does that one not meet your modified requirements? You'll understand, hopefully, that it's difficult to understand what is really meant by a large cone. \$\endgroup\$ – Andy aka Jan 19 '14 at 15:25
  • \$\begingroup\$ Something that radiates in only one side of the sphere. If I cannot a get half sphere (a sphere could also be considered as a cone with 90 degree) I like to get a large cone. The antenna is omnidirectional and it would emit towards analog. I like to avoid rf on analog. \$\endgroup\$ – Ktc Jan 19 '14 at 17:11

I am afraid you won't find an antenna which does not radiate backwards. Even a super-forward-cone-directional Yagi antenna does radiate backwards a little.


What you should do is:

  • use proper PCB layout
  • have a full copper ground plane between your wifi module and analog circuits
  • but remember do not put a ground plane below your chip/patch antenna
  • follow proper RF PCB guidelines - use multiple ground layers and via stitching
  • ensure that the RF module has proper power decoupling
  • it is always a good idea to make a dedicated analog Vdd, preferrably with a linear regulator in order to ensure that power noise (caused not necessarily an RF interference - but rather a digital one, for example processing a received wifi data packet with WEP2) can not hit your analog circuits.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.